Re version 1.05 of 15 October 2001:
Line 375: "allows for the final destination of the message to be an
intermediary MSH other than the 'receiving MSH'": I think the word
"final" should be struck, since the intermediary isn't the "final"
destination.
Line 380: talks about an "ordered-set of discrete parties". I don't
understand what this is talking about. Since when do we route
messages to a set of parties?
Line 469: "There are two logical MIME parts": I think this wording
could be confusing, since it seems to suggest that there are exactly
two MIME parts. I'm not sure exactly what "logical" means here.
Perhaps the intent was to say that there are two KINDS of MIME part,
the Header Container kind and the Payload Container kind. But I think
it would be better to just replace the whole line with something like
"The Message Package contains MIME parts:".
Line 496 uses the term "root part" as if it had already been defined;
in fact it has never been used up to this point. It sort of gets
defined on line 507. In my opinion, we have the term "Header
Container" and we don't need another synonymous term for the same
thing. Can we say "The fist MIME part of the Message Package is
referred to..."?
Line 910: Refers to section 11.1, which doesn't seem to be right.
Perhaps 7.4.5?
Lines 929-930, section 3.1.7.2: this isn't right. duplicateElimination
being true means *either* at-most-once or once-and-only-once semantics
depending on whether ack/retry is happening. duplicateelimination
being false means *either* best-effort or at-least-once semantics
depending on whether ack/retry is happening. This is all laid out
clearly in section 11.5's table.
Line 2276, section 11.5: in the row of the table labelled "4"
(duplicate elimination yes, acks no), "Essentially best effort" should
be "at-most-once".
-- Dan